tensorflow/models · error · ValueError

Illegal padding value; must be one of "left", "right" or Non

Error message

Illegal padding value; must be one of "left", "right" or None.

What it means

Error "Illegal padding value; must be one of "left", "right" or None." thrown in tensorflow/models.

Source

Thrown at official/nlp/modeling/layers/kernel_attention.py:73

      be divisible by the chunk_length.

  Returns:
    Padded tensor with shape[axis] divisible by chunk_length.
  """
  if padding is None:
    return tensor
  shape = tf.shape(tensor)
  rank = tf.rank(tensor)
  if axis < 0:
    axis += rank
  axis_length = shape[axis]
  pad_length = -axis_length % chunk_length
  if padding == "right":
    axis_paddings = [[0, pad_length]]
  elif padding == "left":
    axis_paddings = [[pad_length, 0]]
  else:
    raise ValueError(
        "Illegal padding value; must be one of \"left\", \"right\" or None.")
  paddings = tf.concat([
      tf.zeros([axis, 2], dtype=tf.int32), axis_paddings,
      tf.zeros([rank - axis - 1, 2], dtype=tf.int32)
  ],
                       axis=0)
  return tf.pad(tensor, paddings)


def split_tensor_into_chunks(tensor, axis, chunk_length):
  """Reshape tensor along given axis using chunk_length.

  Args:
    tensor: Input tensor.
    axis: Reshape tensor along this axis.
    chunk_length: Split the axis into [axis/chunk_length, chunk_length]

  Returns:

View on GitHub (pinned to e006f5f0d5)

When it happens

Trigger: Thrown at official/nlp/modeling/layers/kernel_attention.py:73 when the library encounters an invalid state.

Common situations: See trigger scenarios.

AI-assisted analysis of tensorflow/models@e006f5f0d5 (2026-08-24). Data as JSON: /api/errors/b2127ca5da2d2037. Report an issue: GitHub.